Jump table optimization #3
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
cme
reviewed
Feb 14, 2023
cme
reviewed
Feb 14, 2023
A custom lowering method for BR_JT ISDOpcode has been added. We're emitting a MipsISD::BR_JT node with new pseudo instruction LoadJumpTableOffset as a destination. For now, we're setting the entry size to 1B for every jump table leaving the external assembler to decide whether it can be done. We're also emitting relocation directive for that purpose. Each jump table starts with a .jumptable directive.
milica-lazarevic
force-pushed
the
jump-table-opt
branch
2 times, most recently
from
efcca81
to
9747030
Compare
djtodoro
reviewed
Mar 20, 2023
NanoMipsCompressJumpTables optimization pass should decide if there's room for optimization. Before pass runs, the entry size for each table is set to 4B.
An architecture-specific option mno-jump-table-opt has been added. When set, the command line argument mno-jump-table-opt disables jump table optimization. Therefore, table entries are fixed-size.
milica-lazarevic
force-pushed
the
jump-table-opt
branch
from
9747030
to
53c6c58
Compare
|
Can we set the target branch for this to nanomips-llvm13-submission so we can merge it there for interal testing, please? (And we'll then merge to nanomips-llvm13) |
milica-lazarevic
changed the base branch from
nanomips-llvm13
to
nanomips-llvm13-submission
cme
pushed a commit
that referenced
this pull request
Sep 18, 2023
We experienced some deadlocks when we used multiple threads for logging using `scan-builds` intercept-build tool when we used multiple threads by e.g. logging `make -j16` ``` (gdb) bt #0 0x00007f2bb3aff110 in __lll_lock_wait () from /lib/x86_64-linux-gnu/libpthread.so.0 #1 0x00007f2bb3af70a3 in pthread_mutex_lock () from /lib/x86_64-linux-gnu/libpthread.so.0 #2 0x00007f2bb3d152e4 in ?? () #3 0x00007ffcc5f0cc80 in ?? () #4 0x00007f2bb3d2bf5b in ?? () from /lib64/ld-linux-x86-64.so.2 #5 0x00007f2bb3b5da27 in ?? () from /lib/x86_64-linux-gnu/libc.so.6 #6 0x00007f2bb3b5dbe0 in exit () from /lib/x86_64-linux-gnu/libc.so.6 #7 0x00007f2bb3d144ee in ?? () #8 0x746e692f706d742f in ?? () #9 0x692d747065637265 in ?? () #10 0x2f653631326b3034 in ?? () #11 0x646d632e35353532 in ?? () #12 0x0000000000000000 in ?? () ``` I think the gcc's exit call caused the injected `libear.so` to be unloaded by the `ld`, which in turn called the `void on_unload() __attribute__((destructor))`. That tried to acquire an already locked mutex which was left locked in the `bear_report_call()` call, that probably encountered some error and returned early when it forgot to unlock the mutex. All of these are speculation since from the backtrace I could not verify if frames 2 and 3 are in fact corresponding to the `libear.so` module. But I think it's a fairly safe bet. So, hereby I'm releasing the held mutex on *all paths*, even if some failure happens. PS: I would use lock_guards, but it's C. Reviewed-by: NoQ Differential Revision: https://reviews.llvm.org/D118439 (cherry picked from commit d919d02)
cme
pushed a commit
that referenced
this pull request
Sep 18, 2023
Change https://reviews.llvm.org/D140059 exposed the following crash in Z3Solver, where bit widths were not checked consistently with that change. This change makes the check consistent, and fixes the crash. ``` clang: <root>/llvm/include/llvm/ADT/APSInt.h:99: int64_t llvm::APSInt::getExtValue() const: Assertion `isRepresentableByInt64() && "Too many bits for int64_t"' failed. ... Stack dump: 0. Program arguments: clang -cc1 -internal-isystem <root>/lib/clang/16/include -nostdsysteminc -analyze -analyzer-checker=core,unix.Malloc,debug.ExprInspection -analyzer-config crosscheck-with-z3=true -verify reproducer.c #0 0x00000000045b3476 llvm::sys::PrintStackTrace(llvm::raw_ostream&, int) <root>/llvm/lib/Support/Unix/Signals.inc:567:22 #1 0x00000000045b3862 PrintStackTraceSignalHandler(void*) <root>/llvm/lib/Support/Unix/Signals.inc:641:1 #2 0x00000000045b14a5 llvm::sys::RunSignalHandlers() <root>/llvm/lib/Support/Signals.cpp:104:20 #3 0x00000000045b2eb4 SignalHandler(int) <root>/llvm/lib/Support/Unix/Signals.inc:412:1 ... #9 0x0000000004be2eb3 llvm::APSInt::getExtValue() const <root>/llvm/include/llvm/ADT/APSInt.h:99:5 <root>/llvm/lib/Support/Z3Solver.cpp:740:53 clang::ASTContext&, clang::ento::SymExpr const*, llvm::APSInt const&, llvm::APSInt const&, bool) <root>/clang/include/clang/StaticAnalyzer/Core/PathSensitive/SMTConv.h:552:61 ``` Reviewed By: steakhal Differential Revision: https://reviews.llvm.org/D142627 (cherry picked from commit f027dd5)
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Currently, LLVM for nanomips uses jumptables with 4-byte entries that are absolute addresses of target labels. This patchset resolves it in the following way. First, we're emitting 1-byte entries unconditionally. Then, we're introducing a new MIR pass that checks whether it is okay to compress jump table entries. The third patch introduces the mno-jump-table-opt option that enables reverting to absolute jump tables.